The Use of PROXYL Nitroxides in Nitroxide-Mediated Polymerization

Chapter 32

Downloaded by UNIV OF TEXAS AT AUSTIN on March 29, 2018 | https://pubs.acs.org Publication Date: June 26, 2003 | doi: 10.1021/bk-2003-0854.ch032

The Use of PROXYL Nitroxides in Nitroxide-Mediated Polymerization Neil R. Cameron, Catherine A . Bacon, and Alistair J. Reid Department of Chemistry, University of Durham, South Road, Durham DH1 3LE, United Kingdom

The use of alkoxyamines derived from substituted P R O X Y L nitroxides is studied and performance is compared to analogous species obtained from TEMPO. PROXYLs are found to have significant differences relative to TEMPO: more rapid styrene polymerization; the ability to bring about the living polymerization of n-butyl acrylate; and a lower propensity to undergo disproportionation. The latter is suggested to be the key parameter producing the different behaviour of P R O X Y L nitroxides.

452

© 2003 American Chemical Society

Matyjaszewski; Advances in Controlled/Living Radical Polymerization ACS Symposium Series; American Chemical Society: Washington, DC, 2003.

453

+

,0-N

•0-N

R


Figure J. Equilibrium between dormant and active chains in NMP. Nitroxide-mediated polymerization (NMP) is one of three main controlled radical polymerization (CRP) techniques(7). At the heart of N M P is the equilibrium between dormant and active centres Sci. up when radical polymerizations are conducted in the presence of sufficient amounts of nitroxide (Figure 1). NMP has been the subject of intense study for several years, following initial work by Rizzardo, Solomon, Moad et al.(2) Subsequent work carried out by Georges and eoworkers(3,4) employing TEMPO (species on right hand side of Figure 1) as the mediator indicated that this species had limitations, most seriously an inability to polymerize monomers other than styrenes. More recent work has produced a number of acyclic nitroxides that have much greater monomer scope than TEMPO(5,6). Differences between nitroxides have been shown to be due to a fine balance between kd, kc and kp for the monomer in question(7). Our work has focused on PROXYLs, which are the 5-membered ringAnal.oguesof TEMPO. Derivatives of these substituted adjacent to the ring Ν atom are much more accessible synthetically than TEMPO Anal.ogues, allowing us to investigate the influence of structure on activity. Other groups have also studied PROXYLs and importantly have observed differences in behaviour compared to TEMPO. Veregin et ai.(4) demonstrated that 3-carboxy P R O X Y L mediated styrene polymerization more rapidly than TEMPO. A 2,5diphenyl-subsituted derivative was also found to result in a significantly faster polymerization of styrene than TEMPO (8). Hawker, Braslau et al. also employed this mediator and similarly observed a faster rate of styrene polymerization compared to TEMPO(tf). Furthermore, w-butyl acrylate could be polymerized in its presence, although broad polydispersities (~ 2) were obtained. Yamada and coworkers(P) found that ring-substituted P R O X Y L species led to rate enhancements over TEMPO or P R O X Y L itself. Here we describe our studies into the use of alkoxyamines derived from substituted PROXYLs in NMP.

Experimental Toluene (solvent for ESR studies) was distilled before use. TEMPO was purified by vacuum sublimation. w-Butyl acrylate and styrene were freed of inhibitor by passing through basic alumina and were distilled under N immediately prior to use. THF and diethyl ether were purified by heating at reflux over and distilling from Na/benzophenone under N . C H C N was heated at reflux over and distilled from CaH and stored over 4Â molecular seives. A l l other chemicals and solvents were used as received from commercial suppliers 2

2

3

2


454 (mainly Aldrich). Size exclusion chromatography (SEC) with refractive index, viscosity and light scattering detectors, using THF as a solvent, was employed to determine molecular weights and molecular weight distributions. ESR spectra were recorded on a Bruker E M X spectrometer fitted with a variable temperature probe. NMR spectra were obtained with a Varian Innova 400 fitted with a variable temperature probe, operating at either 400 MHz (*H) or 100 M H z ( C), using either CDC1 or d -toluene as the solvent and with tetramethylsilane (TMS) as an internal standard. Mass spectroscopy was performed with a Micromass Auto Spec. The nitroxides and alkoxyamines employed were prepared according to procedures described elsewhere(70,//). The species used in the present work are shown in Figure 2. Di-ter/-butyl peroxalate (DTBPO) was prepared by a literature method(72) (Caution: DTBPO when dry is known to be shock sensitive and to detonate in contact with metal objects). A typical procedure for polymerizations is as follows: a solution of monomer (39 mmol.) and alkoxyamine (0.125 mmol.) in a round-bottomed flask was degassed by 3 freeze-pump-thaw cycles. The flask was back-filled with N and divided between several previously purged GC vials with PTFE seals, by syringe transfer. The vials were heated at 125 °C in an oil bath, withdrawn periodically and immediately cooled in an ice bath. Conversion was determined by H N M R spectroscopy, molecular weights and polydispersity were determined without further purification by SEC. 13


3

8

2

!

Results and Discussion

Nitroxide and Alkoxyamine Synthesis Nitroxides 1 to 4 were generally obtained in low yields (16 - 36%), however usable quantities were produced and the starting materials are cheap. Alkoxyamines 5 to 9 were prepared by trapping of the product of addition of tert-butoxy radicals to styrene with the appropriate nitroxide. For these reactions, yields were higher (typically 70 - 80%) apart from the synthesis of 9, where it is presumed that 4 and DTBPO, the source of terf-butoxy radicals, participate in a redox reaction(ZJ). 10 and 11 were prepared from the corresponding nitroxide and styrene in the presence of Jacobsen's catalyst(/4).

Styrene Polymerization Alkoxyamines 5 to 9 were tested for their ability to mediate the polymerization of styrene. The kinetic plots from heating styrene in their presence at 125°C are shown in Figure 3.



Figure 2. Nitroxides and alkoxyamines used in the present work


in

456 1.8 τ 1.6 1.4 -

•

•

Ά

H

0.8 -

•

•

Χ •

•

χ


ϊ 0.60.4 0.2 t 0 0

2

4

6

8

Time/h Figure 3. Kinetic plots for the po lymerization ofstyrene at 125 °C with: 5 (·); 6 (m); 7 ( L); 8 (4); and 9 (X). (Adapted with permissionfromreferenced 1). Copyright2000 Wiley-VCH). The alkoxyamines can be seen to fall broadly into one of two classes: those that are 'TEMPO-like' (5, 6 and 9); and those that mediate styrene polymerization at an appreciably higher rate (7 and 8). Polymerization with 7 is in fact around 2.5 times faster than with 5. The cause of these differences could be either steric or electronic (or both). Comparing 5 with 7 suggests that steric effects are important, as has been found by other gTO\ips(5 6,15,16). The similarity in polymerization rate between 7 and 8 implies that an electron withdrawing group does not increase the rate of polymerization further, however an electron donating (by resonance) species (9) appears to counteract the steric effect of the phenyl substituent and decrease the polymerization rate to that observed in the presence of the TEMPO-derived species. Thus, the relationship between sterics and electronics and how these influence polymerization rate is a complex one. It should be pointed out that these results conflict somewhat with semiempirical molecular orbital calculations of Moad and Rizzardo, who predicted that nitroxides with electron donating subsituents would result in alkoxyamines with lower C-0 bond homolysis energies(/6). Molecular weight data for polymerizations conducted in the presence of 5 to 9 are depicted in Figure 4. For all apart from 8, M„ is seen to grow linearly with conversion and there is good agreement with predicted values (dotted line). This suggests that the polymerization of styrene is well controlled in the presence of P R O X Y L derived alkoxyamines. The evolution of polydispersity with conversion was also studied. In all cases, polydispersity decreases with time to around 1.3, in agreement with the Persistent Radical Effect(77). The molecular weight distribution for the polymerization in the presence of 8 is broader than the others (1.6 at high conversion), which agrees with the conclusion that this alkoxyamine leads to poorer control. t


457


η-Butyl Acrylate Polymerization Since alkoxyamine 7 was found to lead to a more rapid polymerization of styrene than the species derived from TEMPO (5), the polymerization of w-butyl acrylate was attempted. Our initial studies involved heating a solution of 5 in nbutyl acrylate at 125 °C for 12 h., which produced very little conversion. In contrast, 7 under the same conditions led to 96 % conversion after only 4 h. The polymerization with 7 was repeated in the presence of small quantities of added nitroxide 2; the results are presented in Figure 5. An excess nitroxide concentration of 0.10 equivalents leads to an induction period of around 1 hour, whereas lower added quantities of 2 result in shorter induction periods. In each case, once started the polymerization proceeds at a constant rate with a linear increase in ln([M]o/[M]). The reason for the induction period is unknown, but it suggests that the position of the equilibrium is changing during the early stages of the polymerization. This could be due to nitroxide decomposition(M), however we have no reason to believe that our PROXYL nitroxides are unstable on the timescale of polymerization. An alternative explanation is that there is a source of external radicals, e.g. from trace levels of impurities capable of initiation such as peroxides. The molecular weights of the resulting poiy(«-butyl acrylate)s were investigated by SEC. M was found to increase very rapidly to around 50,000 at low conversion and then did not rise much on increasing conversion. On the other hand, M and especially M increase to very high levels with increasing conversion, as does polydispersity (Figure 6). These data indicate that control over the polymerization mediated by 7 is poor. However, it is also probable that branching is occurring, which is well known for w-butyl acrylate(ZP) and has n

w

z

m

-ι

50 φ

40 -

•

ο 30 20 •

v

-

ΑΧ

·

10 020

40

60

80

100

Conversion(% ) Figure 4. M„ against conversion for the polymerization of styrene in the presence of alkoxyamines (key as in Figure 3). (Adapted with permission from reference^ I). Copyright 2000 Wiley- VCH)


458

2.5

•

Φ

2

Ι

1.5

1

1


Β

0.5

I • • *

0 20

40

60

Time/min

Figure 5. Kinetic plot of the polymerization of η-butyl acrylate at 125°C in the presence of excess quantities of 2: 0.03 eq. (+); 0.05 eq. (M); 0.10 eq. ( A).

0

10 20 30 40 50 60 70 80 90 100 Conversion (%)

Figure 6. Molecular weight and polydispersity against conversion for the polymerization of η-butyl acrylate in the presence of 7 and 0.05 eq. of 2: M„ (·); M (M); M (k);MJM (X) w

z

n


459 been observed during both its atom transfer radical polymerization(20) and N M P in bulk and miniemulsion(27). Quantitative C NMR spectroscopy was used to demonstrate that branching was indeed occurring in our polymerizations (data not shown). Despite the poor control over the polymerization of w-butyl acrylate by 7, we decided to investigate the ability of the resulting polymer to act as a macroinitiator for subsequent polymerizations. The polymerization of a charge of styrene was seen to proceed linearly, indicating that the PB A macroinitiator is indeed capable of initiation. The molecular weight data of the resulting p(BA-S) copolymers were obtained, again by SEC. Interestingly, M was seen to increase with conversion, and the polydispersity to decrease (Figure 7). 13


n

Ο

80 -τ

2.25

70 60

2

50 40 30 -

h 1.75 1.5

20 10

1.25 1

020

40

60

Conversion (% ) Figure 7. Molecular weight and polydispersity data for the polymerization of styrene initiated by a PBA macroinitiator: M„ (*); MJM (·).. n

This indicates a controlled polymerization, suggesting that the P B A macroinitiator is capable of producing a block copolymer and that the polymerization of styrene proceeds in a controlled manner. SEC data obtained with a light scattering detector indicated that the molar mass of the P B A macroinitiator increases on reinitiation with little or no evidence of bimodality in the block copolymer trace, showing that the macroinitiator contained little or no dead P B A material (F igure 8). The ability of PBA to initiate and induce some control in the polymerization of styrene with little evidence of dead material indicates that the vast majority of PBA chains possess a nitroxide moiety. However, the polymerizations are characterized by poor control and broad polydispersities. Fischer(7,77) has pointed out that living character and polymerization control are separate phenomena and that it is indeed possible to have a living polymerization, producing little or no dead material, that gives a broad polydispersity product. Livingness is determined by Κ (see Figure 1), which has a limiting value, dependent on k and k for the monomer in question, above which large fractions of dead material are produced at high conversion. On the other hand, control is p

t


460 determined by the product kdkc, which should be larger than a threshold value, again peculiar to each monomer, required to give a low polydispersity. Thus, for a givenSci.of values of k , k and initial initiator concentration there is a range of values of kd and kc that gives rise to a polymerization that is both living and controlled. It may be the case that the product kdkc for 7 in w-butyl acrylate is relatively small, leading to poor control, while K(=kd/kc) is still sufficiently low to ensure that little dead material is produced. The differences in performance between 7 and TEMPO could also be due to differences in kd and k . 7 has been found to give a higher rate of styrene polymerization than 5, which suggests a higher value of K . However, it is known that TEMPO is prone to causing hydrogen abstraction(22-24) and this may in fact be the crucial difference between 7 and 5. p

t


c

Retention volume/ml Figure 8. SEC traces for a poly(n-butyl acrylate-b-styrene) copolymer and the originalpoly(n-butyl acrylate) macroinitiator (right-hand trace).

Alkoxyamine Homolysis and Disproportionation Studying the fundamental reactions of alkoxyamines, including homolysis and disproportionation, is key to understanding their behaviour in N M P . Therefore, we were keen to investigate the properties of our alkoxyamines to elucidate the reasons for their different performances. The most direct method of studying alkoxyamine homolysis is quantitative electron spin resonance (ESR). ESR spectra were collected from alkoxyamines


461 5, 6, 7, 10 and 11 cleaved in the presence of oxygen as a radical scavenger at different temperatures(25). A linear fit to the double integral with time data allows the calculation of alkoxyamine bond homolysis rate constants (kd), which leads to an Arrhenius plot (ln(kd) vs. 1/T) for each alkoxyamine system. In each case, a nonlinear least squares fitting algorithm was used to obtain the Arrhenius parameters, namely activation energy E and the pre-exponential factor A , from the slope and intercept respectively. The Arrhenius parameters for 5, 6, 7, 10 and 11 are shown in Table 1. It can be seen that the alkoxyamines fall into two classes. Those with 1-phenethyi residues (10 and 11) have significantly higher values of E than alkoxyamines derived from ter/-butoxy radicals (5 to 7). The influence of the nitroxide fragment on alkoxyamine homolysis appears less significant. For the series 5 to 7 there is hardly any effect and indeed E appears to increase slightly on changing from a TEMPO to a substituted P R O X Y L substituent. With the 1phenethyl derived alkoxyamines there is a drop in E when the nitroxide residue is varied and this result is in line with our expectations based on previous work. a


a

a

a

Table 1. Arrhenius Parameters for Alkoxyamines Investigated Alkoxyamine A/s 1 3.2xl0 6 2.0xl0 7 2.05xl0 10 3.23xl0 11 1.61xl0

10

9

n

14

13

EJjkJmoT ) 102.5 104.3 109.8 129.9 121.8

Fischer et al.(2d) performed quantitative ESR measurements on a series of 27 alkoxyamines based on six nitroxides with different carbon centred radical fragments. It was suggested that the observed variation of rate constants was not due to changes in Arrhenius frequency factor but was the result of bond strength differences. The frequency factors observed were similar to those obtained previously by Scaiano and coworkers(22); the majority of systems evaluated had an average value of 2.6 χ 10 s* with a 2.5 fold variation about this average. The difficulty in discussing variations in frequency factor is a consequence of a statistical enthalpy-entropy compensation effect(27) caused by measurement and fitting errors, which masks any real effect when the variation is small. For this reason it was difficult to draw conclusions from values of the frequency factor, as no general trend with nitroxide or transient fragments was evident and reported values vary. The larger variation in activation energies (which more or 14

1


462 less equates to bond dissociation energy) was more reliably assigned to chemical differences (mainly steric). Homolysis activation energies do not explain the observed difference in behaviour of the alkoxyamines described here (E for 5 is lower than that for 7). In an attempt to rationalize this, we examined their decomposition via hydrogen abstraction to generate an unsaturated species and hydroxylamine (Figure 9).


a

Figure 9. Disproportionation of alkoxyamine 10.

Heating 10 at 125°C for a period of time resulted in the formation of significant quantities styrene, as determined by H N M R spectroscopy (spectrum not shown). The method of Fukuda et d\.(24) was used to determine the disproportionation rate constant for 10, giving a value of k ^ = 4.8 χ 10" s" . This compares with the value of 4.5 χ 10* s" determined by Fukuda and coworkers for the same alkoxyamine at 140°C. In contrast, heating related alkoxyamine 5 for 16 hours resulted in the production of very small quantities of vinyl containing species at a level that was too low to obtain a reliable integration (spectrum not shown). This indicates that the introduction of a β substituent greatly reduces the propensity for TEMPO-derived alkoxyamines to undergo disproportionation, in agreement with findings by other workers(22,2

The Use of PROXYL Nitroxides in Nitroxide-Mediated Polymerization

Recommend Documents